Blog
The OCaml Planet RSS
Articles and videos contributed by both experts, companies and passionate developers from the OCaml community. From in-depth technical articles, project highlights, community news, or insights into Open Source projects, the OCaml Planet RSS feed aggregator has something for everyone.
Want your Blog Posts or Videos to Show Here?
To contribute a blog post, or add your RSS feed, check out the Contributing Guide on GitHub.
The OCaml community participated in the December 2025 round of Outreachy internships. We had four interns working on Raven, OCaml-TIFF and YOCaml. This meeting was an opportunity for the interns to present their work and for the community to as...
Reworking the TESSERA Zarr store layout after community feedback, Springer's API woes for evidence synthesis, vibecoding introspection, and git remote helpers for ATProto.
Tessera produces global land-cover embeddings at 0.1-degree resolution, roughly 11 km square at the equator. For each year and each grid tile, there is a directory containing NumPy files of the embeddings.
We’ve all been using Claude via the prompt, and some have even ventured into running claude --dangerously-skip-permissions in a nice sandbox like avsm/claude-ocaml-devcontainer.
This post walks through how to use mtelvers/day10 to compare which opam packages build successfully under two different compiler configurations.
Evidence synthesis at the DEFRA science conference, TESSERA transcoding and building a new SPA, OpenStreetMap/DuckDB bindings in OxCaml, and early thoughts on vibecoding etiquette.
For Pi Day, I have implemented the same algorithm in both OCaml and OxCaml and compared the generated assembly and runtime performance.
How we restructured TESSERA's geospatial embeddings from millions of individual numpy files into sharded Zarr v3 stores for efficient HTTP streaming, enabling everything from single-pixel mobile lookups to regional-scale analysis with just a couple of range requests.
Following my previous CPU vs GPU post I started thinking about what the ONNX inference engine actually did and if it could be replicated in OxCaml with SIMD.
